import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
import missingno as msno
import plotly.express as px

sns.set_style('whitegrid')
pd.set_option('display.max_columns', None)
import warnings
warnings.simplefilter('ignore')

df = pd.read_csv(r'D:\Datasets\survey.csv')

df.head()

	Timestamp	Age	Gender	Country	state	self_employed	family_history	treatment	work_interfere	no_employees	remote_work	tech_company	benefits	care_options	wellness_program	seek_help	anonymity	leave	mental_health_consequence	phys_health_consequence	coworkers	supervisor	mental_health_interview	phys_health_interview	mental_vs_physical	obs_consequence	comments
0	2014-08-27 11:29:31	37	Female	United States	IL	NaN	No	Yes	Often	6-25	No	Yes	Yes	Not sure	No	Yes	Yes	Somewhat easy	No	No	Some of them	Yes	No	Maybe	Yes	No	NaN
1	2014-08-27 11:29:37	44	M	United States	IN	NaN	No	No	Rarely	More than 1000	No	No	Don't know	No	Don't know	Don't know	Don't know	Don't know	Maybe	No	No	No	No	No	Don't know	No	NaN
2	2014-08-27 11:29:44	32	Male	Canada	NaN	NaN	No	No	Rarely	6-25	No	Yes	No	No	No	No	Don't know	Somewhat difficult	No	No	Yes	Yes	Yes	Yes	No	No	NaN
3	2014-08-27 11:29:46	31	Male	United Kingdom	NaN	NaN	Yes	Yes	Often	26-100	No	Yes	No	Yes	No	No	No	Somewhat difficult	Yes	Yes	Some of them	No	Maybe	Maybe	No	Yes	NaN
4	2014-08-27 11:30:22	31	Male	United States	TX	NaN	No	No	Never	100-500	Yes	Yes	Yes	No	Don't know	Don't know	Don't know	Don't know	No	No	Some of them	Yes	Yes	Yes	Don't know	No	NaN

df.shape

(1259, 27)

df.tail()

	Timestamp	Age	Gender	Country	state	self_employed	family_history	treatment	work_interfere	no_employees	remote_work	tech_company	benefits	care_options	wellness_program	seek_help	anonymity	leave	mental_health_consequence	phys_health_consequence	coworkers	supervisor	mental_health_interview	phys_health_interview	mental_vs_physical	obs_consequence	comments
1254	2015-09-12 11:17:21	26	male	United Kingdom	NaN	No	No	Yes	NaN	26-100	No	Yes	No	No	No	No	Don't know	Somewhat easy	No	No	Some of them	Some of them	No	No	Don't know	No	NaN
1255	2015-09-26 01:07:35	32	Male	United States	IL	No	Yes	Yes	Often	26-100	Yes	Yes	Yes	Yes	No	No	Yes	Somewhat difficult	No	No	Some of them	Yes	No	No	Yes	No	NaN
1256	2015-11-07 12:36:58	34	male	United States	CA	No	Yes	Yes	Sometimes	More than 1000	No	Yes	Yes	Yes	No	No	Don't know	Somewhat difficult	Yes	Yes	No	No	No	No	No	No	NaN
1257	2015-11-30 21:25:06	46	f	United States	NC	No	No	No	NaN	100-500	Yes	Yes	No	Yes	No	No	Don't know	Don't know	Yes	No	No	No	No	No	No	No	NaN
1258	2016-02-01 23:04:31	25	Male	United States	IL	No	Yes	Yes	Sometimes	26-100	No	No	Yes	Yes	No	No	Yes	Don't know	Maybe	No	Some of them	No	No	No	Don't know	No	NaN

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1259 entries, 0 to 1258
Data columns (total 27 columns):
 #   Column                     Non-Null Count  Dtype 
---  ------                     --------------  ----- 
 0   Timestamp                  1259 non-null   object
 1   Age                        1259 non-null   int64 
 2   Gender                     1259 non-null   object
 3   Country                    1259 non-null   object
 4   state                      744 non-null    object
 5   self_employed              1241 non-null   object
 6   family_history             1259 non-null   object
 7   treatment                  1259 non-null   object
 8   work_interfere             995 non-null    object
 9   no_employees               1259 non-null   object
 10  remote_work                1259 non-null   object
 11  tech_company               1259 non-null   object
 12  benefits                   1259 non-null   object
 13  care_options               1259 non-null   object
 14  wellness_program           1259 non-null   object
 15  seek_help                  1259 non-null   object
 16  anonymity                  1259 non-null   object
 17  leave                      1259 non-null   object
 18  mental_health_consequence  1259 non-null   object
 19  phys_health_consequence    1259 non-null   object
 20  coworkers                  1259 non-null   object
 21  supervisor                 1259 non-null   object
 22  mental_health_interview    1259 non-null   object
 23  phys_health_interview      1259 non-null   object
 24  mental_vs_physical         1259 non-null   object
 25  obs_consequence            1259 non-null   object
 26  comments                   164 non-null    object
dtypes: int64(1), object(26)
memory usage: 265.7+ KB

Cleansing

 df.columns = df.columns.str.lower()

round(df.isna().sum()/df.shape[0]*100,2)

timestamp                     0.00
age                           0.00
gender                        0.00
country                       0.00
state                        40.91
self_employed                 1.43
family_history                0.00
treatment                     0.00
work_interfere               20.97
no_employees                  0.00
remote_work                   0.00
tech_company                  0.00
benefits                      0.00
care_options                  0.00
wellness_program              0.00
seek_help                     0.00
anonymity                     0.00
leave                         0.00
mental_health_consequence     0.00
phys_health_consequence       0.00
coworkers                     0.00
supervisor                    0.00
mental_health_interview       0.00
phys_health_interview         0.00
mental_vs_physical            0.00
obs_consequence               0.00
comments                     86.97
dtype: float64

df['country'].value_counts()

United States             751
United Kingdom            185
Canada                     72
Germany                    45
Ireland                    27
Netherlands                27
Australia                  21
France                     13
India                      10
New Zealand                 8
Poland                      7
Switzerland                 7
Sweden                      7
Italy                       7
South Africa                6
Belgium                     6
Brazil                      6
Israel                      5
Singapore                   4
Bulgaria                    4
Austria                     3
Finland                     3
Mexico                      3
Russia                      3
Denmark                     2
Greece                      2
Colombia                    2
Croatia                     2
Portugal                    2
Moldova                     1
Georgia                     1
Bahamas, The                1
China                       1
Thailand                    1
Czech Republic              1
Norway                      1
Romania                     1
Nigeria                     1
Japan                       1
Hungary                     1
Bosnia and Herzegovina      1
Uruguay                     1
Spain                       1
Zimbabwe                    1
Latvia                      1
Costa Rica                  1
Slovenia                    1
Philippines                 1
Name: country, dtype: int64

df['state'].unique()

array(['IL', 'IN', nan, 'TX', 'TN', 'MI', 'OH', 'CA', 'CT', 'MD', 'NY',
       'NC', 'MA', 'IA', 'PA', 'WA', 'WI', 'UT', 'NM', 'OR', 'FL', 'MN',
       'MO', 'AZ', 'CO', 'GA', 'DC', 'NE', 'WV', 'OK', 'KS', 'VA', 'NH',
       'KY', 'AL', 'NV', 'NJ', 'SC', 'VT', 'SD', 'ID', 'MS', 'RI', 'WY',
       'LA', 'ME'], dtype=object)

df = df.drop(['country','state','timestamp','comments'], axis=1)

round(df.describe(include = 'all'),2)

	age	gender	self_employed	family_history	treatment	work_interfere	no_employees	remote_work	tech_company	benefits	care_options	wellness_program	seek_help	anonymity	leave	mental_health_consequence	phys_health_consequence	coworkers	supervisor	mental_health_interview	phys_health_interview	mental_vs_physical	obs_consequence
count	1.259000e+03	1259	1241	1259	1259	995	1259	1259	1259	1259	1259	1259	1259	1259	1259	1259	1259	1259	1259	1259	1259	1259	1259
unique	NaN	49	2	2	2	4	6	2	2	3	3	3	3	3	5	3	3	3	3	3	3	3	2
top	NaN	Male	No	No	Yes	Sometimes	6-25	No	Yes	Yes	No	No	No	Don't know	Don't know	No	No	Some of them	Yes	No	Maybe	Don't know	No
freq	NaN	615	1095	767	637	465	290	883	1031	477	501	842	646	819	563	490	925	774	516	1008	557	576	1075
mean	7.942815e+07	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
std	2.818299e+09	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
min	-1.726000e+03	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
25%	2.700000e+01	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
50%	3.100000e+01	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
75%	3.600000e+01	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
max	1.000000e+11	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN

df['gender'].value_counts().reset_index()

	index	gender
0	Male	615
1	male	206
2	Female	121
3	M	116
4	female	62
5	F	38
6	m	34
7	f	15
8	Make	4
9	Male	3
10	Woman	3
11	Cis Male	2
12	Man	2
13	Female (trans)	2
14	Female	2
15	Trans woman	1
16	msle	1
17	male leaning androgynous	1
18	Neuter	1
19	cis male	1
20	queer	1
21	Female (cis)	1
22	Mail	1
23	cis-female/femme	1
24	A little about you	1
25	Malr	1
26	p	1
27	femail	1
28	Cis Man	1
29	Guy (-ish) ^_^	1
30	Enby	1
31	Agender	1
32	Androgyne	1
33	Male-ish	1
34	maile	1
35	Trans-female	1
36	Cis Female	1
37	something kinda male?	1
38	Mal	1
39	Male (CIS)	1
40	queer/she/they	1
41	non-binary	1
42	Femake	1
43	woman	1
44	Nah	1
45	All	1
46	fluid	1
47	Genderqueer	1
48	ostensibly male, unsure what that really means	1

df['gender'].replace(['Male ', 'male', 'M', 'm', 'Male', 'Cis Male',
                     'Man', 'cis male', 'Mail', 'Male-ish', 'Male (CIS)',
                      'Cis Man', 'msle', 'Malr', 'Mal', 'maile', 'Make',], 'Male', inplace = True)

df['gender'].replace(['Female ', 'female', 'F', 'f', 'Woman', 'Female',
                     'femail', 'Cis Female', 'cis-female/femme', 'Femake', 'Female (cis)',
                     'woman',], 'Female', inplace = True)
df["gender"].replace(['Female (trans)', 'queer/she/they', 'non-binary',
                     'fluid', 'queer', 'Androgyne', 'Trans-female', 'male leaning androgynous',
                      'Agender', 'A little about you', 'Nah', 'All',
                      'ostensibly male, unsure what that really means',
                      'Genderqueer', 'Enby', 'p', 'Neuter', 'something kinda male?',
                      'Guy (-ish) ^_^', 'Trans woman',], 'Other', inplace = True)

df['gender'].value_counts()

Male      991
Female    247
Other      21
Name: gender, dtype: int64

df.loc[df.age<12, 'age'] =15
df.loc[df.age>75, 'age'] = 75

df[df['age'] > 80].head()

	age	gender	self_employed	family_history	treatment	work_interfere	no_employees	remote_work	tech_company	benefits	care_options	wellness_program	seek_help	anonymity	leave	mental_health_consequence	phys_health_consequence	coworkers	supervisor	mental_health_interview	phys_health_interview	mental_vs_physical	obs_consequence

df['age'].hist()

<AxesSubplot:>

msno.matrix(df)

<AxesSubplot:>

from sklearn.model_selection import train_test_split
train_data, test_data = train_test_split(df,test_size=0.15,stratify=df['treatment'],random_state=42 )

print(f'Train Data Dimensions: {train_data.shape}\n Test data dimensions:{test_data.shape}')

Train Data Dimensions: (1070, 23)
 Test data dimensions:(189, 23)

EDA

health = train_data.copy()

health.columns

Index(['age', 'gender', 'self_employed', 'family_history', 'treatment',
       'work_interfere', 'no_employees', 'remote_work', 'tech_company',
       'benefits', 'care_options', 'wellness_program', 'seek_help',
       'anonymity', 'leave', 'mental_health_consequence',
       'phys_health_consequence', 'coworkers', 'supervisor',
       'mental_health_interview', 'phys_health_interview',
       'mental_vs_physical', 'obs_consequence'],
      dtype='object')

plt.figure(figsize=(8,8))
px.histogram(health, x ='treatment', color ='treatment')

<Figure size 576x576 with 0 Axes>

px.histogram(health, x ='age', color = 'treatment')

#need to remove age12< entries as we perveive it to be false entries

px.histogram(health, x ='self_employed', color = 'treatment', barmode = 'group')

sns.countplot(data =health, x ='work_interfere')

<AxesSubplot:xlabel='work_interfere', ylabel='count'>

health['work_interfere'].value_counts(normalize = True)

Sometimes    0.468009
Never        0.210900
Rarely       0.174171
Often        0.146919
Name: work_interfere, dtype: float64

px.histogram(health, x='work_interfere', color ='treatment', barmode ='group')

#family history if mental illnesss

px.histogram(health, x= 'family_history', color ='treatment', barmode = 'group')

#remote work hours

sns.countplot(data = health, x ='remote_work')

<AxesSubplot:xlabel='remote_work', ylabel='count'>

px.histogram(health, x ='remote_work', color ='treatment',barmode ='group')

sns.countplot(data=health, x ='benefits')

<AxesSubplot:xlabel='benefits', ylabel='count'>

health['benefits'].value_counts(normalize = True)

Yes           0.377570
Don't know    0.324299
No            0.298131
Name: benefits, dtype: float64

px.histogram(health, x = 'benefits',color='treatment',barmode='group')

sns.countplot(data = health , x = 'care_options')

<AxesSubplot:xlabel='care_options', ylabel='count'>

health['care_options'].value_counts(normalize=True)

No          0.394393
Yes         0.357009
Not sure    0.248598
Name: care_options, dtype: float64

px.histogram(health, x = 'care_options',color='treatment',barmode='group')

#how many employees does your company have?

sns.countplot(data = health , x = 'no_employees')

<AxesSubplot:xlabel='no_employees', ylabel='count'>

px.histogram(health, x = 'no_employees',color='benefits',barmode='group')

px.histogram(health, x = 'no_employees',color='treatment',barmode='group')

#Has your employer ever discussed mental health as part of an employee wellness program

sns.countplot(data = health , x = 'wellness_program')

<AxesSubplot:xlabel='wellness_program', ylabel='count'>

px.histogram(health, x = 'wellness_program',color='treatment',barmode='group')

#Does your employer provide resources to learn more about mental health issues and how to seek help?¶

sns.countplot(data = health , x = 'seek_help')

<AxesSubplot:xlabel='seek_help', ylabel='count'>

px.histogram(health, x = 'seek_help',color='treatment',barmode='group')

#Is your anonymity protected if you choose to take advantage of mental health or substance abuse treatment resources?¶

sns.countplot(data = health , x = 'anonymity')

<AxesSubplot:xlabel='anonymity', ylabel='count'>

px.histogram(health, x = 'anonymity',color='treatment',barmode='group')

#How easy is it for you to take medical leave for a mental health condition?¶

sns.countplot(data = health , x = 'leave')
plt.xticks(rotation=45)
plt.show()

px.histogram(health, x = 'leave',color='treatment',barmode='group')

#Do you think that discussing a mental health issue with your employer would have negative consequences?¶

sns.countplot(data = health , x = 'mental_health_consequence' )

<AxesSubplot:xlabel='mental_health_consequence', ylabel='count'>

px.histogram(health, x = 'mental_health_consequence',color='treatment',barmode='group')

sns.countplot(data = health , x = 'coworkers' )

<AxesSubplot:xlabel='coworkers', ylabel='count'>

px.histogram(health, x = 'coworkers',color='treatment',barmode='group')

#Do you feel that your employer takes mental health as seriously as physical health?¶

sns.countplot(data = health , x = 'mental_vs_physical' )

<AxesSubplot:xlabel='mental_vs_physical', ylabel='count'>

px.histogram(health, x = 'mental_vs_physical',color='treatment',barmode='group')

#Have you heard of or observed negative consequences for coworkers with mental health conditions in your workplace?¶

sns.countplot(data = health , x = 'obs_consequence')

<AxesSubplot:xlabel='obs_consequence', ylabel='count'>

Starting with we need to treat missing values for the columns 'work_interfere' and 'self_employed'

se_mode = train_data['self_employed'].mode().values[0]
train_data['self_employed'].fillna(se_mode,inplace = True)
train_data[train_data['work_interfere'].isna()]['treatment'].value_counts()

No     223
Yes      3
Name: treatment, dtype: int64

A 99% of the people who didn't answet the question 'Whether mental health interfers with work' don't seek help. If we think null values as a category, 'NEVER' of all categories of column_work_interfere.

train_data['work_interfere'].fillna('Never',inplace= True)

msno.bar(train_data, color = (0.5,0.2,0.35))

<AxesSubplot:>

As there are no missing values we proceed to the next step.

X_train = train_data.drop('treatment',axis=1)
Y_train = train_data['treatment'].copy()

All columns are categorical, except age.

train_data.columns

Index(['age', 'gender', 'self_employed', 'family_history', 'treatment',
       'work_interfere', 'no_employees', 'remote_work', 'tech_company',
       'benefits', 'care_options', 'wellness_program', 'seek_help',
       'anonymity', 'leave', 'mental_health_consequence',
       'phys_health_consequence', 'coworkers', 'supervisor',
       'mental_health_interview', 'phys_health_interview',
       'mental_vs_physical', 'obs_consequence'],
      dtype='object')

Cardinal Information

gender_cols = ['Female','Male','Other']
self_employed_cols = ['No','Yes']
family_history_cols = ['No','Yes']
work_interfere_cols = ['Never','Rarely','Sometimes','Often']
no_employees_cols = ['1-5','6-25','26-100','100-500','500-1000','More than 1000']
remote_work_cols = ['No','Yes']
tech_company_cols = ['No','Yes']
benefits_cols = ['No','Don\'t know','Yes'] 
care_options_cols = ['No','Not sure','Yes']
wellness_program_cols  =['No','Don\'t know','Yes']
seek_help_cols = ['No','Don\'t know','Yes']
anonymity_cols = ['No','Don\'t know','Yes']
leave_cols = [ 'Very easy', 'Somewhat easy',"Don't know" ,'Somewhat difficult','Very difficult']
mental_health_consequence_cols = ['No','Maybe','Yes']
phys_health_consequence_cols = ['No','Maybe','Yes']
coworkers_col = ['No','Some of them','Yes']
supervisor_cols = ['No','Some of them','Yes']
mental_health_interview_cols = ['No','Maybe','Yes']
phys_health_interview_cols = ['No','Maybe','Yes']
mental_vs_physical_cols = ["Don't know",'No','Yes']
obs_consequence_cols = ['No','Yes']

columns_for_encoder = [gender_cols,self_employed_cols,family_history_cols,work_interfere_cols,no_employees_cols,remote_work_cols,
                            tech_company_cols,benefits_cols,care_options_cols,wellness_program_cols,seek_help_cols,anonymity_cols,leave_cols,
                            mental_health_consequence_cols,phys_health_consequence_cols,coworkers_col,supervisor_cols,mental_health_interview_cols,
                            phys_health_interview_cols,mental_vs_physical_cols,obs_consequence_cols]

features = list(X_train.columns)

from sklearn.preprocessing import OrdinalEncoder
ord_encoder = OrdinalEncoder(categories = list(columns_for_encoder))
X_train[features[1:]] = ord_encoder.fit_transform(X_train.iloc[:,1:])

X_train.head()

	age	gender	self_employed	family_history	work_interfere	no_employees	remote_work	tech_company	benefits	care_options	wellness_program	seek_help	anonymity	leave	mental_health_consequence	phys_health_consequence	coworkers	supervisor	mental_health_interview	phys_health_interview	mental_vs_physical	obs_consequence
676	22	1.0	0.0	0.0	0.0	2.0	0.0	1.0	0.0	0.0	0.0	0.0	1.0	2.0	1.0	1.0	1.0	2.0	1.0	1.0	0.0	0.0
88	29	0.0	0.0	0.0	2.0	2.0	0.0	1.0	2.0	2.0	0.0	0.0	1.0	2.0	1.0	0.0	1.0	1.0	0.0	1.0	1.0	1.0
86	39	1.0	1.0	0.0	3.0	1.0	0.0	1.0	0.0	0.0	0.0	0.0	1.0	4.0	1.0	0.0	1.0	0.0	0.0	1.0	1.0	1.0
1210	24	1.0	0.0	1.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	1.0	2.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0
505	46	1.0	0.0	0.0	2.0	5.0	0.0	1.0	2.0	2.0	2.0	2.0	1.0	2.0	1.0	0.0	1.0	0.0	0.0	0.0	2.0	0.0

from sklearn.preprocessing import StandardScaler
std_scaler = StandardScaler()
X_train[features] = std_scaler.fit_transform(X_train)

X_train

	age	gender	self_employed	family_history	work_interfere	no_employees	remote_work	tech_company	benefits	care_options	wellness_program	seek_help	anonymity	leave	mental_health_consequence	phys_health_consequence	coworkers	supervisor	mental_health_interview	phys_health_interview	mental_vs_physical	obs_consequence
676	-1.363472	0.419109	-0.362047	-0.803837	-1.135675	-0.243796	-0.659027	0.480937	-1.319345	-1.111531	-0.650242	-0.869401	-0.461953	0.267568	0.223297	1.212902	0.040848	1.062155	1.546143	0.357235	-0.963586	-0.406912
88	-0.409844	-1.904448	-0.362047	-0.803837	0.720967	-0.243796	-0.659027	0.480937	1.125156	1.197863	-0.650242	-0.869401	-0.461953	0.267568	0.223297	-0.569797	0.040848	-0.124177	-0.461591	0.357235	0.233902	2.457534
86	0.952483	0.419109	2.762070	-0.803837	1.649288	-0.831323	-0.659027	0.480937	-1.319345	-1.111531	-0.650242	-0.869401	-0.461953	2.108713	0.223297	-0.569797	0.040848	-1.310508	-0.461591	0.357235	0.233902	2.457534
1210	-1.091007	0.419109	-0.362047	1.244033	-1.135675	-1.418850	-0.659027	0.480937	-1.319345	-1.111531	-0.650242	-0.869401	-0.461953	0.267568	0.223297	1.212902	-1.577948	-1.310508	-0.461591	-1.058475	-0.963586	-0.406912
505	1.906111	0.419109	-0.362047	-0.803837	0.720967	1.518785	-0.659027	0.480937	1.125156	1.197863	1.893665	1.700376	-0.461953	0.267568	0.223297	-0.569797	0.040848	-1.310508	-0.461591	-1.058475	1.431390	-0.406912
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
988	-1.227240	0.419109	-0.362047	1.244033	0.720967	-0.243796	-0.659027	0.480937	1.125156	0.043166	1.893665	1.700376	1.396280	-0.653004	-1.082318	-0.569797	-1.577948	-1.310508	-0.461591	0.357235	-0.963586	-0.406912
598	1.497413	0.419109	-0.362047	-0.803837	-1.135675	0.931258	1.517389	0.480937	-0.097095	0.043166	0.621712	0.415487	-0.461953	0.267568	-1.082318	-0.569797	1.659645	1.062155	-0.461591	-1.058475	1.431390	-0.406912
363	-0.001146	0.419109	-0.362047	1.244033	-0.207354	1.518785	-0.659027	-2.079275	1.125156	1.197863	1.893665	1.700376	1.396280	0.267568	-1.082318	-0.569797	0.040848	1.062155	-0.461591	1.772946	1.431390	-0.406912
690	-1.227240	0.419109	-0.362047	-0.803837	0.720967	-0.831323	-0.659027	0.480937	1.125156	-1.111531	-0.650242	-0.869401	-0.461953	-0.653004	0.223297	1.212902	0.040848	-0.124177	-0.461591	-1.058475	-0.963586	-0.406912
191	-0.546076	0.419109	-0.362047	1.244033	0.720967	-0.243796	-0.659027	0.480937	1.125156	-1.111531	-0.650242	-0.869401	-0.461953	-0.653004	0.223297	1.212902	0.040848	-1.310508	-0.461591	0.357235	0.233902	-0.406912

1070 rows × 22 columns

from sklearn.preprocessing import LabelEncoder
lb_encoder = LabelEncoder()
Y_train = lb_encoder.fit_transform(Y_train)

Data Preprocessing part ends here. Further, we move to Model Building.

Model Building and evaluation

from sklearn.metrics import f1_score
from sklearn.model_selection import cross_val_score

def train_evaluate(model,X_train,Y_train,name):
    model.fit(X_train,Y_train)
    y_pred = model.predict(X_train)
    f1_train = f1_score(Y_train,y_pred)

    #Cross validation
    f1_val = cross_val_score(model,X_train,Y_train,scoring='f1',cv=10)
    
    # returning the scores
    score = pd.DataFrame({'Name' : name ,'F1_score_trainset' : [f1_train], 'F1_score_validationset' : [f1_val.mean()]})
    return score

Logistic Regression

from sklearn.linear_model import LogisticRegression
log_reg = LogisticRegression(penalty='l1',solver='liblinear')
train_evaluate(log_reg,X_train,Y_train,'Logistic Regression')

	Name	F1_score_trainset	F1_score_validationset
0	Logistic Regression	0.826521	0.817808

from sklearn.tree import DecisionTreeClassifier
dt_clf = DecisionTreeClassifier(max_leaf_nodes=4,random_state=42)
train_evaluate(dt_clf,X_train,Y_train,'DecisionTreeClassifier')

	Name	F1_score_trainset	F1_score_validationset
0	DecisionTreeClassifier	0.851698	0.851769

from sklearn.svm import SVC
svc_clf = SVC()
train_evaluate(svc_clf, X_train, Y_train, 'Support Vector Classifier')

	Name	F1_score_trainset	F1_score_validationset
0	Support Vector Classifier	0.885159	0.834448

from sklearn.ensemble import RandomForestClassifier
rnd_clf = RandomForestClassifier(random_state=42)
train_evaluate(rnd_clf, X_train,Y_train, 'Random Forest Classifier')

	Name	F1_score_trainset	F1_score_validationset
0	Random Forest Classifier	1.0	0.842216

from sklearn.ensemble import AdaBoostClassifier
dt_clf_ada = DecisionTreeClassifier()
Ada_clf = AdaBoostClassifier(base_estimator = dt_clf_ada,random_state=42)
train_evaluate(Ada_clf,X_train,Y_train, "Ada Boost Classifier")

	Name	F1_score_trainset	F1_score_validationset
0	Ada Boost Classifier	1.0	0.767202

from sklearn.ensemble import GradientBoostingClassifier
gdb_clf = GradientBoostingClassifier(random_state=42,subsample=0.8)
train_evaluate(gdb_clf,X_train,Y_train, "Gradient Boosting Classifier")

	Name	F1_score_trainset	F1_score_validationset
0	Gradient Boosting Classifier	0.897596	0.850009

from xgboost import XGBClassifier
xbg_clf = XGBClassifier(verbosity=0)
train_evaluate(xbg_clf,X_train,Y_train, "XGBoost Classifier")

	Name	F1_score_trainset	F1_score_validationset
0	XGBoost Classifier	1.0	0.828707

Among the models, we narrow down the most promising models to fine tune the parameters further

from sklearn.model_selection import GridSearchCV
param_distribs = {
        'kernel': ['linear', 'rbf','polynomial'],
        'C': [0.01,0.01,0.1,0.15,0.2,0.25,0.5,0.75,1,2,10,100],
        'gamma': [1, 0.1, 0.01, 0.001, 0.0001],
    }
svm_clf = SVC()
grid_cv = GridSearchCV(svm_clf , param_grid = param_distribs,
                              cv=5,scoring='f1',
                              verbose=1)
grid_cv.fit(X_train,Y_train)

grid_cv.best_estimator_

SVC(C=0.5, gamma=0.01)

train_evaluate(grid_cv.best_estimator_,X_train,y_train,"SVC Tuned")

---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-133-7cf7c05c6769> in <module>
----> 1 train_evaluate(grid_cv.best_estimator_,X_train,y_train,"SVC Tuned")

NameError: name 'y_train' is not defined